Recent work demonstrates that veterans exhibit higher suicide risk compared with the general U.S. population. Despite progress in understanding risk factors for suicidal behavior, the pathogenesis is poorly understood, including alterations in the neural circuitry underlying affective instability (AI) associated with suicidal behavior. AI is a trait that cuts across multiple psychiatric disorders in a dimensional manner. Two core components of AI are emotional reactivity and regulation. In addition to examining brain activity and functional connectivity in key neural circuitry underlying these two components of AI in veterans at low and high risk for suicide, this project examines affective startle modulation, a translational, psychophysiological measure mediated by the amygdala that provides a reliable, low-cost, nonverbal metric of AI components. Progress in the prevention and prediction of suicidal behavior would be facilitated by the identification of quantitative measures of AI-related neural activity/connectivity and/or affective startle modulation in response to validated unpleasant pictures and may serve as dimensional psychophysiological endophenotypes of risk for suicide. Additionally, we will examine three reliable self-report measures of AI (measuring lability, intensity, and emotion regulation) which may serve as a dimensional phenotype of suicidal behavior. Examining this combination of neural circuits, physiology, and behavior in veterans at low (non-suicidal psychiatric controls) and high risk (suicidal ideators, suicide attempters) for suicide, as well as healthy controls, holds great promise for understanding the pathogenesis of suicidal behavior, and identifying targets that may ultimately provide novel treatment intervention to reverse such pathogenic processes. The proposed longitudinal study will focus on AI as a critical dimension that is directly associated with risk for suicide and identify the neural-circuitry disturbances that underlie emotion processing abnormalities in veterans at low and high risk for suicide. To do so, we will characterize a sample of 144 veterans, 36 in each of the four groups. All participants will receive rigorous diagnostic and clinical assessments (including several well-validated measures of AI and suicide risk), and undergo 3T functional MRI and affective startle modulation measurement while they perform passive emotion-processing/reactivity and active emotion-regulation tasks. At 6-month follow-up, all participants will repeat the startle assessment in order to examine test-retest reliability. Clinical assessment follow-up will be done at 6- and 12-months in the three patient groups. The project aims to identify behavioral, neurobiological, and psychophysiological features underlying suicidal behavior in veterans and determine whether baseline psychophysiological measures predict suicidal behavior at 12-month follow- up. Impact: Our prospective, multi-modal design promises to help uncover the mechanisms by which biological and psychological factors give rise to suicidal behavior. The proposed research may aid in prospectively identifying veterans at greatest risk for suicidal behavior.